diff options
Diffstat (limited to 'results/classifier/no-thinking-deepseek-r1:70b/reasoning/instruction/1793119')
| -rw-r--r-- | results/classifier/no-thinking-deepseek-r1:70b/reasoning/instruction/1793119 | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/results/classifier/no-thinking-deepseek-r1:70b/reasoning/instruction/1793119 b/results/classifier/no-thinking-deepseek-r1:70b/reasoning/instruction/1793119 new file mode 100644 index 00000000..4aed4f63 --- /dev/null +++ b/results/classifier/no-thinking-deepseek-r1:70b/reasoning/instruction/1793119 @@ -0,0 +1,3 @@ +The issue described involves incorrect behavior of floating-point operations when emulating AArch64 on QEMU compared to real hardware. The problem specifically occurs with the `fdiv` instruction when FPCR is set to zero. This points to a flaw in how QEMU handles the emulation of the `fdiv` instruction, which falls under the category of an instruction-level issue. + +instruction \ No newline at end of file |